But they purposefully use CLS in Search to increase clicks on Ads https://twitter.com/andyhattemer/status/1262564268890820609
You present this as a fact, but it would be absurd that Google would use such a cheap and easily detected trick to increase CTR. It would be bordering on ad fraud and I'm sure that Google, of all companies, knows better than that.
Occam's Razor says that this is a stupid async content loading bug, which they subsequently fixed. I've never seen this happen and when I just tried it without adblocker with that exact search term, it didn't - the page loaded with the ad.
3 years ago and I wouldn't believed it at all but around 2 years ago I saw it happen consistently with a colleague at the desk next to me.
I cannot say for sure that it wasn't an extension in his browser but I can say for sure that I think Google has been really busy tearing down the mountains of trust they had before 2007 - 2009.
For this to not be an accident, one would have to assume that Google actually makes more money from those invalid clicks, and that someone decided that yep, rendering ads asynchronously was a decent and legal approach at increasing advertising revenue, and requested the GMail team to implement it.
This kind of corporate misbehavior is not unheard of, but I just can't imagine it happening at Google.
It's much more likely that this is just unfortunate UX design to "improve" rendering performance without considering users on slow connections.
(I can reproduce this one just fine in desktop GMail - on the first render of the "Promotions" tab, the ads render asynchronously)
[1]: https://support.google.com/google-ads/answer/42995?hl=en
[2]: https://www.blog.google/products/ads/preventing-accidental-c...
'Unfortunate UX design to 'improve' rendering' is the plausible-deniability they can use to justify this.
> This kind of corporate misbehavior is not unheard of, but I just can't imagine it happening at Google.
I definitely can, I don't think anywhere is immune to this once you reach a certain scale. They have a profit-motive to make money, they will absolutely try and get away with as much as they possibly can.
And the re-ordering happens as your mails and the ads are loading! You might be about to tap your email, then the ads load in and you suddenly click on an ad. Or you want to tap the top row, but the app decides to put a different email above the ads and you end up tapping into the wrong mail because it was reordered just before the tap.
No, they don't. This is false. It's a mechanism called Media Engagement Index, Google properties have zero advantage, and any site can get a high score.
Chrome ships with a preloaded MEI assembled from global telemetry data, which is then trained locally:
https://www.chromium.org/audio-video/autoplay/autoplay-pre-s...
Would they have made the same choice of preloading a default seed if they had no properties in the seed ? who knows
Once they reached a dominant ad network position their whole strategy has been “advancing the web is advancing our revenue”, and it bled into mobile to the point where building and maintaining a whole ecosystem for free makes sense as long as they stay the search and ad engine of choice (that’s the only thing they’ll fight to impose).
Chrome is built in the same optics: push forward the web and webapps as long as search is theirs.
There are other examples where only the large sites benefit while everybody else has to play by stricter rules: "EU Parliament bans geoblocking, exempts Netflix and other streaming services" -- https://www.dw.com/en/eu-parliament-bans-geoblocking-exempts...
EDIT: User teraflop posted a link to the list of "sites that are allowed to autoplay video even without any prior media engagement" right here in this thread https://news.ycombinator.com/item?id=24818178
My guess from someone who had to develop a web video player at work, many websites will attempt to autoplay the video with sound and if it fails, it's easy to catch the failure event, they will mute the video and try again.
Web browsers are also capable of determining that autoplay on technically-not-load-but-automatic counts as autoplay. (There's even text in the spec about it.) In particular, they can tell whether it is in response to a user action/gesture on the site or not.
https://developers.google.com/web/updates/2017/09/autoplay-p...
The preloaded list is in the source code (https://github.com/chromium/chromium/blob/master/chrome/brow...) but it's encoded as a finite state automaton that makes it a bit difficult to enumerate the list of whitelisted domains.
Here is the code: https://gist.github.com/NeatMonster/e9cdb01441a3cd842e6a20fd...
And here is the plain-text list: https://gist.github.com/NeatMonster/e9cdb01441a3cd842e6a20fd...
[0] https://github.com/chromium/chromium/blob/615d5eed47c10d8890...
https://www.chromium.org/audio-video/autoplay/autoplay-pre-s...
According to the MEI it actively measures user behavior and one of the most important measures is that a video is unmuted. From the document:
“The MEI is meant to allow media heavy websites (e.g. YouTube, Netflix) that rely on autoplay for their core experience. It is a non-goal to allow websites with a “good media behaviour” to autoplay without restrictions”
It doesn’t sound too good, and still doesn’t really explain how everything is seeded.
The preimage space is finite and easily enumerated.
These are the kind of tricks a shady company would do. So disappointed what Google is doing to the web the last few years.
It isn't obvious to me from this that Google are privileging their own sites above others here